A Probabilistic Model for JSON Data

JSON is an exchange data format for applications nowadays. To manage JSON data efficiently, it is important to use a proper model for JSON data. A straightforward model for JSON is a tree-structured model, in which nodes are types of JSON data, and edges are values of the nodes. As uncertainties are common and natural in real environments, it is realistic for a JSON model to deal with uncertainties of JSON data. This paper first analyzes the state-of-art uncertainty in data management fields, including relational database, XML data, and graph data, etc. Then based on the above observations, a probabilistic JSON model is proposed based on the tree model of JSON data is proposed, in which probability settings are incorporated in the (key, value) pairs of JSON model. Finally, JSONPath is extended to deal with probability computing of JSON data queries.